Why is improving the diagnosis of tuberculosis important?
Every year, an estimated one million children and young adolescents become ill with tuberculosis, and around 226,000 die from the disease. Tuberculosis is caused by the bacterium Mycobacterium tuberculosis and mostly affects the lungs (pulmonary tuberculosis), though it can affect other sites in the body (extrapulmonary tuberculosis). Signs and symptoms of pulmonary tuberculosis include cough, fever, night sweats, and weight loss. Signs and symptoms of extrapulmonary tuberculosis depend on the site of disease. When detected early and treated effectively, tuberculosis is largely curable.
Not recognizing tuberculosis (false negative) early may result in delayed diagnosis and treatment, severe illness, and death. An incorrect tuberculosis diagnosis (false positive) may result in anxiety, unnecessary treatment (which can involve medication side effects), and the possibility of missing alternative diagnoses which warrant treatment.
What was the aim of this review?
To determine the accuracy of Xpert Ultra in children with symptoms of tuberculosis for diagnosing pulmonary tuberculosis, tuberculous meningitis (affecting membranes that surround the brain and spinal cord), lymph node tuberculosis (a painful swelling of one or more lymph nodes, which are bean-shaped structures that help fight infection), and rifampicin resistance.
What did this review study?
Xpert Ultra, a World Health Organization-recommended rapid test that simultaneously detects tuberculosis and rifampicin resistance in adults and children with tuberculosis symptoms. Rifampicin is an important medicine used to treat tuberculosis. For tuberculosis diagnosis, we assessed results against two different benchmarks: tuberculosis culture (a method used to grow bacteria on nutrient-rich media) and a composite definition based on symptoms, chest X-ray, sputum microscopy (examination under a microscope of mucus and other matter coughed up from the lungs), and culture. For rifampicin resistance detection, we assessed results against drug susceptibility testing or line probe assay (a rapid laboratory-based test for detecting tuberculosis bacteria).
What were the main results in this review?
We included 14 studies. For pulmonary tuberculosis, we analysed 335 data sets (around 26,000 participants). No studies evaluated Xpert Ultra accuracy for tuberculous meningitis or lymph node tuberculosis. Three studies evaluated Xpert Ultra accuracy for detection of rifampicin resistance.
For a population of 1000 children:
• where 100 have pulmonary tuberculosis in sputum according to culture results:
- 101 would be Xpert Ultra-positive, and of these, 26 (26%) would not have pulmonary tuberculosis (false positive); and
- 899 would be Xpert Ultra-negative, and of these, 25 (3%) would have tuberculosis (false negative).
• where 100 have pulmonary tuberculosis in gastric aspirate (collection of lung and oral secretions from the stomach) according to culture results:
- 97 would be Xpert Ultra-positive, and of these, 27 (28%) would not have pulmonary tuberculosis (false positive); and
- 903 would be Xpert Ultra-negative, and of these, 30 (3%) would have tuberculosis (false negative).
• where 100 have pulmonary tuberculosis in stool according to culture results:
- 74 would be Xpert Ultra-positive and of these, 18 (24%) would not have pulmonary tuberculosis (false positive); and
- 926 would be Xpert Ultra-negative, and of these, 44 (5%) would have tuberculosis (false negative).
• where 100 have pulmonary tuberculosis in nasopharyngeal aspirate (secretions from the uppermost part of the throat, behind the nose) according to culture results:
- 66 would be Xpert Ultra-positive, and of these, 22 (33%) would not have pulmonary tuberculosis (false positive); and
- 934 would be Xpert Ultra-negative, and of these, 56 (6%) would have tuberculosis (false negative).
Xpert Ultra accurately detected rifampicin resistance, but there were few studies and only three children with rifampicin resistance included.
How confident are we in the results of this review?
For pulmonary tuberculosis, we are fairly confident because we included studies from different countries and used two different benchmarks, though neither is perfect. However, the evidence base is still limited and there were few studies with few children for one of the specimen types (nasopharyngeal aspirate).
For rifampicin resistance, we identified few studies with very few children with rifampicin resistance, so we are less confident.
What children do the results of this review apply to?
Children and young adolescents (birth to 14 years) who are HIV-positive or HIV-negative, with signs or symptoms of pulmonary tuberculosis. The results also apply to children with severe pneumonia or malnutrition and tuberculosis symptoms. In this review, we did not identify any studies that evaluated Xpert Ultra accuracy for tuberculous meningitis or lymph node tuberculosis.
What are the implications of this review?
The results suggest that Xpert Ultra in sputum, gastric aspirate, stool, and nasopharyngeal aspirate is an accurate method for detecting pulmonary tuberculosis and rifampicin resistance in children.
Using Xpert Ultra in sputum, gastric aspirate, stool, and nasopharyngeal aspirate, the risk of missing a diagnosis of pulmonary tuberculosis (confirmed by culture) is low, suggesting that only a small number of children will not receive treatment. The risk of incorrectly diagnosing a child as having pulmonary tuberculosis is slightly higher. This may result in some children receiving unnecessary treatment.
How up to date is this review?
This review updates our previous review and includes evidence published up to 9 March 2021.
We found Xpert Ultra sensitivity to vary by specimen type, with sputum having the highest sensitivity, followed by gastric aspirate and stool. Nasopharyngeal aspirate had the lowest sensitivity. Xpert Ultra specificity was high against both microbiological and composite reference standards. However, the evidence base is still limited, and findings may be imprecise and vary by study setting. Although we found Xpert Ultra accurate for detection of rifampicin resistance, results were based on a very small number of studies that included only three children with rifampicin resistance. Therefore, findings should be interpreted with caution. Our findings provide support for the use of Xpert Ultra as an initial rapid molecular diagnostic in children being evaluated for tuberculosis.
Every year, an estimated one million children and young adolescents become ill with tuberculosis, and around 226,000 of those children die. Xpert MTB/RIF Ultra (Xpert Ultra) is a molecular World Health Organization (WHO)-recommended rapid diagnostic test that simultaneously detects Mycobacterium tuberculosis complex and rifampicin resistance. We previously published a Cochrane Review 'Xpert MTB/RIF and Xpert MTB/RIF Ultra assays for tuberculosis disease and rifampicin resistance in children'. The current review updates evidence on the diagnostic accuracy of Xpert Ultra in children presumed to have tuberculosis disease. Parts of this review update informed the 2022 WHO updated guidance on management of tuberculosis in children and adolescents.
To assess the diagnostic accuracy of Xpert Ultra for detecting: pulmonary tuberculosis, tuberculous meningitis, lymph node tuberculosis, and rifampicin resistance, in children with presumed tuberculosis.
Secondary objectives
To investigate potential sources of heterogeneity in accuracy estimates. For detection of tuberculosis, we considered age, comorbidity (HIV, severe pneumonia, and severe malnutrition), and specimen type as potential sources.
To summarize the frequency of Xpert Ultra trace results.
We searched the Cochrane Infectious Diseases Group Specialized Register, MEDLINE, Embase, three other databases, and three trial registers without language restrictions to 9 March 2021.
Cross-sectional and cohort studies and randomized trials that evaluated Xpert Ultra in HIV-positive and HIV-negative children under 15 years of age. We included ongoing studies that helped us address the review objectives. We included studies evaluating sputum, gastric, stool, or nasopharyngeal specimens (pulmonary tuberculosis), cerebrospinal fluid (tuberculous meningitis), and fine needle aspirate or surgical biopsy tissue (lymph node tuberculosis). For detecting tuberculosis, reference standards were microbiological (culture) or composite reference standard; for stool, we also included Xpert Ultra performed on a routine respiratory specimen. For detecting rifampicin resistance, reference standards were drug susceptibility testing or MTBDRplus.
Two review authors independently extracted data and, using QUADAS-2, assessed methodological quality judging risk of bias separately for each target condition and reference standard. For each target condition, we used the bivariate model to estimate summary sensitivity and specificity with 95% confidence intervals (CIs). We stratified all analyses by type of reference standard. We summarized the frequency of Xpert Ultra trace results; trace represents detection of a very low quantity of Mycobacterium tuberculosis DNA. We assessed certainty of evidence using GRADE.
We identified 14 studies (11 new studies since the previous review). For detection of pulmonary tuberculosis, 335 data sets (25,937 participants) were available for analysis. We did not identify any studies that evaluated Xpert Ultra accuracy for tuberculous meningitis or lymph node tuberculosis. Three studies evaluated Xpert Ultra for detection of rifampicin resistance. Ten studies (71%) took place in countries with a high tuberculosis burden based on WHO classification. Overall, risk of bias was low.
Detection of pulmonary tuberculosis
Sputum, 5 studies
Xpert Ultra summary sensitivity verified by culture was 75.3% (95% CI 64.3 to 83.8; 127 participants; high-certainty evidence), and specificity was 97.1% (95% CI 94.7 to 98.5; 1054 participants; high-certainty evidence).
Gastric aspirate, 7 studies
Xpert Ultra summary sensitivity verified by culture was 70.4% (95% CI 53.9 to 82.9; 120 participants; moderate-certainty evidence), and specificity was 94.1% (95% CI 84.8 to 97.8; 870 participants; moderate-certainty evidence).
Stool, 6 studies
Xpert Ultra summary sensitivity verified by culture was 56.1% (95% CI 39.1 to 71.7; 200 participants; moderate-certainty evidence), and specificity was 98.0% (95% CI 93.3 to 99.4; 1232 participants; high certainty-evidence).
Nasopharyngeal aspirate, 4 studies
Xpert Ultra summary sensitivity verified by culture was 43.7% (95% CI 26.7 to 62.2; 46 participants; very low-certainty evidence), and specificity was 97.5% (95% CI 93.6 to 99.0; 489 participants; high-certainty evidence).
Xpert Ultra sensitivity was lower against a composite than a culture reference standard for all specimen types other than nasopharyngeal aspirate, while specificity was similar against both reference standards.
Interpretation of results
In theory, for a population of 1000 children:
• where 100 have pulmonary tuberculosis in sputum (by culture):
- 101 would be Xpert Ultra-positive, and of these, 26 (26%) would not have pulmonary tuberculosis (false positive); and
- 899 would be Xpert Ultra-negative, and of these, 25 (3%) would have tuberculosis (false negative).
• where 100 have pulmonary tuberculosis in gastric aspirate (by culture):
- 123 would be Xpert Ultra-positive, and of these, 53 (43%) would not have pulmonary tuberculosis (false positive); and
- 877 would be Xpert Ultra-negative, and of these, 30 (3%) would have tuberculosis (false negative).
• where 100 have pulmonary tuberculosis in stool (by culture):
- 74 would be Xpert Ultra-positive, and of these, 18 (24%) would not have pulmonary tuberculosis (false positive); and
- 926 would be Xpert Ultra-negative, and of these, 44 (5%) would have tuberculosis (false negative).
• where 100 have pulmonary tuberculosis in nasopharyngeal aspirate (by culture):
- 66 would be Xpert Ultra-positive, and of these, 22 (33%) would not have pulmonary tuberculosis (false positive); and
- 934 would be Xpert Ultra-negative, and of these, 56 (6%) would have tuberculosis (false negative).
Detection of rifampicin resistance
Xpert Ultra sensitivity was 100% (3 studies, 3 participants; very low-certainty evidence), and specificity range was 97% to 100% (3 studies, 128 participants; low-certainty evidence).
Trace results
Xpert Ultra trace results, regarded as positive in children by WHO standards, were common. Xpert Ultra specificity remained high in children, despite the frequency of trace results.